IntelPython / dpnp / 26357963747

78%

Build Type

push

github

Committed by web-flow

Commit Message

Add boolean mask validation when slicing (#2929)

This PR proposes the addition of validation for the shape of masks when
slicing `tensor.usm_ndarray`s and `dpnp_array`s.

This aligns with NumPy for specific edge cases, such as:
```
import dpnp.tensor as dpt

x = dpt.reshape(dpt.arange(3**5, dtype="i4"), (3,) * 5)
ii = dpt.asarray(1)
i0 = dpt.asarray(0, dtype='?')
i1 = dpt.asarray(0, dtype='?')

x[ii, i0[dpt.newaxis], ii, i1[dpt.newaxis], :]
```

which previously would give `usm_ndarray([], dtype=int32)`, but now
raises an error, which aligns with NumPy.

Coverage Stats

1539 of 2874 branches covered (53.55%)

Branch coverage included in aggregate %.

8 of 9 new or added lines in 1 file covered. (88.89%)

1 existing line in 1 file now uncovered.

26431 of 32874 relevant lines covered (80.4%)

7900.4 hits per line